Abstract
Background
Nutrition misinformation spreads rapidly on social media. Young adults particularly rely on influencers and platforms, such as TikTok, for nutrition information. This study examined young adult TikTok users’ trust in nutrition information and nutrition influencers, exploring whether influencer qualifications and message accuracy influence trust.
Methods
We conducted an online experiment using Prolific and Centiment. A total of 1,035 TikTok users aged 18–26 years were recruited from four countries (USA, UK, Australia, Canada), with 1,016 included in the final analysis. Participants were randomly assigned to view one of six TikTok videos manipulating (1) influencer qualification (qualified vs neutral vs not qualified) and (2) information accuracy (accurate vs misinformation). Outcomes were self-reported trust in the influencer and trust in the message content, measured using validated multi-item scales. Differences across conditions were analysed using factorial ANCOVA models in SPSS.
Results
Trust in the influencer and trust in the message differed by qualification condition, with the highest trust observed when the influencer reported a nutrition qualification. Trust in both outcomes was also higher when the video contained accurate information compared with misinformation, with mean differences across conditions ranging from approximately 0.47 to 0.50 points on the scale. Post hoc analyses indicated lower trust in both the influencer and the message among Canadian participants relative to other countries.
Conclusions
Influencer qualifications and information accuracy independently increase young adults’ trust in TikTok nutrition content. These findings support encouraging qualified nutrition professionals to prominently disclose relevant credentials, ensure accuracy in original posts, and actively counter inaccurate nutrition content shared by others.
Keywords: Nutrition, Misinformation, Social media, Trust, TikTok
Introduction
Social media has become a widely used source of health information and, increasingly, health misinformation [1]. While platforms such as Facebook and Twitter have long shaped online health communication, short-form video platforms—particularly TikTok -have rapidly changed how nutrition information is produced and encountered. TikTok’s video-first format and algorithmic content delivery enable health-related content to reach large audiences quickly in engaging, easily digestible forms [2]. The platform surpassed one billion monthly active users in 2021 [3] and is disproportionately used by younger populations with over 60% of young adult smartphone users in the United States report using TikTok [4]. In parallel, the use of social media for nutrition information seeking has grown, particularly among younger users [5]. TikTok is frequently used for food and nutrition inspiration, with 45% of young adults using the platform for recipe inspiration and 70% following influencers [6].
At the same time, nutrition information ecosystems on TikTok may amplify harmful content. A study of nutrition content on TikTok found that weight-centric content (content that frames weight management and achieving a “normal” weight as central to health) and weight loss trends are common and may foster unhealthy eating behaviours [7]. Beyond the promotion of potentially harmful trends, nutrition advice is frequently offered by unqualified sources, and misinformation can spread rapidly in the absence of robust regulation and oversight [7]. More broadly, health-related social media content can be inaccurate, incomplete, outdated, or framed as unverified opinion, rumour, or spam [1, 8]. Absorption of health misinformation may delay legitimate care-seeking and may support harmful dietary practices, including disordered eating, with heightened concern for younger users [7, 9, 10].
Although research on health misinformation on social media has expanded across topics such as vaccines [11] and mental health [2, 12], important gaps remain. In particular, recent reviews call for research examining the accuracy of nutrition information and health misinformation on platforms like TikTok that rely on video, rather than text-based data [13–15]. In addition, highlight the need to examine demographic factors shaping susceptibility to misinformation, as characteristics such as education, socio-economic status, and age can influence individuals’ health literacy, critical evaluation skills, and access to credible information sources. These factors may affect both the ability to identify inaccurate claims and the likelihood of trusting and acting on them. Responding to these calls, this study focuses on nutrition misinformation on TikTok among young adult users.
Influencers are central to contemporary social media information flows and can facilitate the spread of misinformation [16, 17]. pg. 59) define a social media influencer as “…a content generator: one who has a status of expertise in a specific area, who has cultivated a sizable number of captive followers—who are of marketing value to brands—by regularly producing valuable content via social media.” Influencer-generated content operates within electronic word-of-mouth (eWOM) dynamics [18, 19], where users disseminate opinions and experiences, and where high-visibility content creators can shape attitudes and behaviour. Influencers are often perceived as credible within their communities and may be trusted by followers who view their endorsements as honest or authoritative [20]. This trust can be reinforced by platform-visible signals of popularity and endorsement (e.g., follower counts, likes, comments, shares) [21], and strategic self-presentation and branding [21–23].
In nutrition contexts, these dynamics matter because young social media users may struggle to discern credible guidance from misinformation [5, 24, 25]. Even when users value credible sources (e.g., registered dieticians), testimonials and peer influencer content can create perceived legitimacy that reshapes understandings of choices and risks [11, 26]. This makes it important to understand how young adults form trust judgements about both the source (the influencer) and the content (the nutrition message), particularly when encountering unfamiliar influencers.
We draw on Truth Default Theory (TDT) [27] to explain how trust judgements may form under conditions of information overload and limited verification. TDT posits that individuals often default to presuming truthfulness in communication unless prompted to doubt. In algorithmically curated environments, users may rely on cues that reduce cognitive burden and facilitate quick judgements. In the context of nutrition advice, one cue is the perception of appropriate qualifications. When an influencer is perceived as qualified, this may reduce users’ perceived need to continue information searching or information verification, and may reduce dissonance during message processing, thereby increasing trust in both the influencer and the message. Consistent with this cue-based logic, prior research suggests that signals of expertise can increase perceived legitimacy of misinformation in other health contexts [11]. Additionally, influencers’ messages may be perceived as highly credible [28]. Given that young users report higher trust in, and comfort following, credentialed nutrition professionals such as dietitians [5], qualifications may function as a familiar and salient heuristic cue when evaluating unfamiliar influencers. Together, this motivates the expectation that disclosed qualification will increase trust.
H1a
Qualification will have a positive influence on Trust in the Influencer.
H1b
Qualification will have a positive influence on Trust in the Message.
Message accuracy is also likely to shape trust. Users report valuing research-based nutrition advice; for example, 91% of participants in one study demonstrated an understanding of what constitutes evidence-based nutrition information [5], although many still report difficulties identifying accurate information in practice. In COVID-19 misinformation contexts, authentic information was viewed as more credible and associated with higher intentions to engage, and influencer-promoted authentic information was associated with higher credibility and engagement intentions [25, 29]. However, nutrition contexts may present a different challenge: users may have difficulty identifying accurate nutrition information, with 77% demonstrating difficulty in one study [5]. These mixed findings underscore the importance of directly testing whether accuracy shapes trust judgements on TikTok. We therefore hypothesise that accurate nutrition advice will increase trust in both the influencer and the message.
H2a
Information Accuracy will have a positive influence on Trust in the Influencer.
H2b
Information Accuracy will have a positive influence on Trust in the Message.
Methods
We conducted an online between-subjects experiment in which participants were randomly assigned to view one of six TikTok-style videos. The online survey was hosted on the Qualtrics platform and programmed to ensure an equal number of participants were randomly assigned to each video. The design enabled assessment of how (i) influencer qualification and (ii) nutrition information accuracy influenced trust in the influencer and trust in the message. The study received ethical approval from the Research Ethics Committee of Dublin City University, Ireland and all participants provided informed consent prior to participation.
Data collection and participants
The sample size was determined based on feasibility and resource constraints, while ensuring a sufficiently large sample to support experimental comparisons across conditions and detect meaningful differences in trust outcomes. Participants were recruited in December 2023 via the online panel providers Prolific and Centiment, a common approach in social science research [30]. Eligibility criteria were: (a) aged 18–26 years, (b) resident in the United States, United Kingdom, Canada, or Australia, and (c) active TikTok user. These four English-speaking countries were selected to broaden evidence beyond the United States, where much misinformation research is concentrated [31], and because each has substantial TikTok user bases [32].
A total of 1,035 respondents were recruited (USA n = 238; Canada n = 209; UK n = 247; Australia n = 341). Following initial data cleaning, 1,029 participants were retained for descriptive analyses (Table 1). The survey included two attention checks and a manipulation check. Nineteen participants were excluded for failing one or both attention checks, resulting in a final analytic sample of 1,016. Participants were compensated approximately $1.30–$1.50, consistent with panel-provider guidance, and the survey took approximately six minutes to complete.
Table 1.
Recruited sample by country
| Variable | Category | All N = 1029 |
USA N = 238 |
Canada N = 209 |
Australia N = 340 |
UK N = 242 |
|---|---|---|---|---|---|---|
| Gender | Male |
460 (44.7%) |
118 (49.6%) |
84 (40.2%) |
134 (39.4%) |
124 (51.2%) |
| Female |
444 (56%) |
120 (50.4%) |
122 (58.4%) |
196 (57.6%) |
118 (48.8%) |
|
| Non-binary |
9 (0.9%) |
3 (1.4%) |
6 (1.8%) |
|||
| Other/Prefer not to say |
4 (0.4%) |
4 (1.2%) |
||||
| Age | 18–20 years |
234 (22.7%) |
47 (19.7%) |
37 (17.7%) |
101 (29.7%) |
49 (20.2%) |
| 21–23 years |
390 (37.9%) |
88 (37.0%) |
86 (41.1%) |
124 (36.5%) |
92 (38%) |
|
| 24–26 years |
405 (39.4%) |
103 (43.3%) |
86 (41.1%) |
115 (33.8%) |
101 (41.7%) |
|
| Education | High school partially or fully completed |
255 (24.7%) |
44 (18.5%) |
34 (26.4%) |
120 (38.2%) |
47 (19.0%) |
| Some college, no degree |
245 (23.8%) |
67 (28.2%) |
56 (26.8%) |
67 (19.7%) |
55 (22.7%) |
|
| Associates degree or equivalent |
55 (5.3%) |
21 (8.8%) |
13 (6.2%) |
15 (4.4%) |
6 (2.5%) |
|
| Bachelor’s degree |
392 (38.1%) |
88 (37%) |
95 (45.5%) |
105 (30.9%) |
104 (43.0%) |
|
| Master’s degree |
77 (7.5%) |
14 (5.9%) |
11 (5.3%) |
22 (6.5%) |
30 (12.4%) |
|
| PhD or equivalent |
5 (0.5%) |
4 (1.7%) |
- |
1 (0.3%) |
- | |
| TikTok Use | Never |
89 (8.6%) |
18 (7.6%) |
18 (8.6%) |
32 (9.4%) |
21 (8.7%) |
| Sometimes |
280 (27.2%) |
70 (29.4%) |
54 (25.8%) |
87 (25.6%) |
69 (28.5%) |
|
| About half the time |
130 (12.6%) |
31 (13%) |
22 (10.5%) |
51 (15.0%) |
26 (10.7%) |
|
| Most of the time |
165 (16%) |
41 (17.2%) |
41 (19.6%) |
46 (13.5%) |
37 (15.3%) |
|
| Everyday |
365 (35.5%) |
78 (32.8%) |
74 (35.4%) |
124 (36.5%) |
89 (36.8%) |
Stimulus development
This study focused on a nutrition trend commonly observed on social media: influencers recommending high-protein diets exceeding guideline recommendations. Excessive protein intake has been associated with potential long-term kidney harm [33], making this a relevant context for examining trust in nutrition advice.
We recruited a professional actress via a freelancing platform, as most nutrition influencers are female [34]. The actress had a neutral accent and experience in scripted performances and short-form video production was selected. This approach ensured the video stimuli aligned with the style of videos posted by health influencers on TikTok. The use of an actress also ensured that participants had no pre-existing impressions of the influencer.
A script for each experimental condition (six scripts total) was developed by the research team and reviewed by registered nutrition experts. Based on expert feedback, scripts were refined so that accurate conditions aligned with prevailing guidelines irrespective of participants’ country (i.e., World Health Organisation (WHO) guidance), and misinformation conditions clearly diverged from those guidelines. Scripts were modelled on common TikTok nutrition-influencer formats: the influencer began with a frequently asked question, introduced themselves (including qualification information where relevant), and then provided advice. Each script referenced WHO protein intake guidance; in misinformation conditions the influencer explicitly disagreed with the guideline recommendation.
Videos were produced to resemble typical TikTok nutrition advice posts: the influencer spoke directly to camera in neutral clothing against a neutral background, used a handheld microphone, and displayed on-screen captions. The actress recorded and edited the videos iteratively in response to researcher feedback. Each video was approximately one minute in length.
Experimental design and procedure
The experiment used a 3 × 2 factorial design with two manipulations: influencer qualification (qualified; neutral; not qualified) and information accuracy (accurate information; misinformation). Qualification was manipulated as follows: in the qualified condition, the influencer stated they had a nutrition-based qualification; in the neutral condition, the influencer did not mention qualification; and in the not qualified condition, the influencer stated they did not have a nutrition-based qualification. Accuracy was manipulated by aligning advice with WHO guidance (accurate condition) or recommending protein intake substantially above WHO guidance (misinformation condition). The video scripts are included in the Appendix.
The survey consisted of four stages. First, participants reported country of residence, age, and TikTok use. Second, participants were randomly assigned to one video condition and instructed to imagine they were scrolling TikTok when the video appeared. The survey was programmed so the video played in full before participants could proceed. Third, outcomes were assessed using validated measures [28]: trust in the influencer and trust in the message, with both variables measured using 3 items and 5-point Likert scale ranging from 1 = Strongly disagree to 5 = Strongly Agree). The items are outline below in Table 2.
Table 2.
Items for dependent variables
| Dependent variable | Item |
|---|---|
| Trust in the influencer | Based on your impression of the person in the video you just watched, please indicate your level of agreement with each of the following statements |
| 1. I trust this person | |
| 2. I would rely on this person for diet and nutrition advice | |
| 3. I would buy products or services from this person | |
| Trust in the message | Based on your impression of the content in the video you just watched, please indicate your level of agreement with each of the following statements… |
| 1. I trust the information provided in this video | |
| 2. I would rely on the information provided in this video | |
| 3. I would share this information with other people |
Finally, participants reported demographics (gender, education, household income). On completion, participants received a debrief explaining the fictitious nature of the content and the potential inclusion of misinformation; they were instructed not to act on the advice and were provided links to reputable national nutrition guidance relevant to their country of residence.
Data analysis
Testing the experimental design
All data analysis was conducted using SPSS v28 (Statistical Package for the Social Sciences). The first stage of analysis involved testing the experimental design to determine whether the dependent variables varied across the video conditions. Preliminary one-way ANOVAs (Analysis of Variance) assessed overall between-condition variation; hypothesis testing relied on factorial ANCOVAs (Analysis of Covariance) aligned with the 3 × 2 design. As shown in Table 3, the one-way ANOVAs indicate significant between-condition differences for each dependent variable, providing an initial indication that the video conditions produced different trust responses.
Table 3.
One way ANOVA
| Dependent variable | Manipulation | F | P value |
|---|---|---|---|
| Trust in the influencer | Information accuracy | 10.732 | 0.001 |
| Qualification | 7.528 | < 0.001 | |
| Trust in the message | Information accuracy | 11.921 | < 0.001 |
| Qualification | 5.782 | 0.003 |
Hypothesis testing involved a series of ANCOVAs to examine the influence of the experimental manipulations on each dependent variable. ANCOVA allows for univariate generalized linear models (GLM) and is commonly used in experimental designs to precisely assess how manipulations affect dependent variables [35]. Additionally, in this study, nationality was treated as a fixed factor. By controlling for country of residence, we focused on the impact of manipulations on the dependent variables. In post hoc testing, inter-country differences were explored. Given the widespread belief in misinformation across health topics [31], we conducted post hoc analysis to examine the influence of qualification and information accuracy on both dependent variables across each country. The ANCOVAs included country as a factor, first with qualification, followed by information accuracy.
Results
Trust in the influencer
The first dependent variable, trust in the influencer, was examined using ANCOVA to assess how the influencer’s qualification and information accuracy impacted trust. Results indicated that both qualification (F = 7.62, p < 0.001) and information accuracy (F = 10.54, p < 0.01) significantly influenced trust in the influencer. Information accuracy emerged as the stronger driver of trust in the influencer. The interaction between qualification and accuracy was not significant. These findings support H1a and H2a. As shown in Fig. 1 and Table 4, trust in the influencer was highest when the influencer was qualified and shared accurate information, and lowest when the influencer was unqualified and shared misinformation. Age was included as a covariate and was not significant (F = 3.73, p > 0.05).
Fig. 1.
Results: Trust in the Influencer
Table 4.
Descriptive statistics: trust in influencer
| Information accuracy | Qualification | Mean | Std. Deviation | N |
|---|---|---|---|---|
| Accurate information | Qualified | 3.71 | 0.83 | 168 |
| Neutral | 3.50 | 1.02 | 174 | |
| Not qualified | 3.43 | 0.97 | 172 | |
| Total | 3.54 | 0.95 | 514 | |
| Misinformation | Qualified | 3.50 | 0.92 | 176 |
| Neutral | 3.33 | 1.04 | 165 | |
| Not qualified | 3.21 | 0.99 | 174 | |
| Total | 3.35 | 0.98 | 515 | |
| Total | Qualified | 3.60 | 0.88 | 344 |
| Neutral | 3.42 | 1.03 | 339 | |
| Not qualified | 3.32 | 0.98 | 346 | |
| Total | 3.45 | 0.97 | 1029 |
Trust in the message
The second dependent variable, trust in the message, was also analysed using ANCOVA to explore the impact of influencer qualification and information accuracy on trust in the nutrition advice provided. Results indicated that both qualification (F = 5.88, p < 0.01) and information accuracy (F = 11.72, p < 0.001) significantly influenced trust in the message. As with trust in the influencer, information accuracy was the stronger driver. The interaction between qualification and accuracy was not significant. These results support H1b and H2b. As shown in Fig. 2 and Table 5, trust in the message was highest when the influencer was qualified and shared accurate information, and lowest when the influencer was unqualified and shared misinformation. Age was included as a covariate and was not significant (F = 3.18, p > 0.05) .
Fig. 2.
Results: trust in the message
Table 5.
Descriptive statistics: trust in message
| Information accuracy | Qualification | Mean | Std. Deviation | N |
|---|---|---|---|---|
| Accurate information | Qualified | 3.53 | 0.83 | 168 |
| Neutral | 3.33 | 1.00 | 174 | |
| Not qualified | 3.27 | 0.98 | 172 | |
| Total | 3.37 | 0.94 | 514 | |
| Misinformation | Qualified | 3.29 | 0.90 | 176 |
| Neutral | 3.14 | 1.04 | 165 | |
| Not qualified | 3.06 | 0.98 | 174 | |
| Total | 3.17 | 0.98 | 515 | |
| Total | Qualified | 3.41 | 0.87 | 344 |
| Neutral | 3.24 | 1.02 | 339 | |
| Not qualified | 3.16 | 0.98 | 346 | |
| Total | 3.27 | 0.97 | 1029 |
Post hoc testing results
The post hoc test results indicated that qualification (F = 7.53, p < 0.001) and country (F = 6.91, p < 0.001) significantly influenced respondents’ trust in the influencer. Similarly, information accuracy (F = 11.55, p < 0.001) and country (F = 7.19, p < 0.001) significantly influenced respondents’ trust in the influencer.
In the second set of ANCOVAs, qualification (F = 5.91, p < 0.001) and country (F = 5.84, p < 0.001) significantly influenced trust in the message. The final ANCOVA showed that information accuracy (F = 12.46, p < 0.001) and country (F = 6.12, p < 0.001) significantly influenced trust in the message.
The mean scores for both dependent variables by country are summarised in Table 6. For each dependent variable and condition, the country with the highest mean score is highlighted in bold, and the country with the lowest mean score is highlighted in bold italic. Across countries, trust in the influencer tended to be higher when the influencer disclosed a nutrition qualification. In the neutral qualification condition, mean trust in the influencer was lower among Canadian and UK respondents than in the not qualified condition. For trust in the message, mean scores were generally higher when information was accurate; however, Canadian respondents reported comparatively lower trust in both the influencer and the message across conditions. These country-level differences are descriptive and should be interpreted as exploratory.
Table 6.
Post hoc ANCOVA results, mean scores for dependent variables in each country across all six experimental conditions
| Dependent variable | Country | Trust in the influencer | Trust in the message |
|---|---|---|---|
| Qualification | |||
| Qualified | USA | 3.78 | 3.58 |
| Canada | 3.48 | 3.33 | |
| Australia | 3.58 | 3.37 | |
| UK | 3.54 | 3.35 | |
| Neutral | USA | 3.68 | 3.47 |
| Canada | 3.04 | 2.87 | |
| Australia | 3.44 | 3.26 | |
| UK | 3.45 | 3.30 | |
| Not qualified | USA | 3.37 | 3.43 |
| Canada | 3.11 | 3.06 | |
| Australia | 3.26 | 3.24 | |
| UK | 3.52 | 3.27 | |
| Information accuracy | |||
| Accurate information | USA | 3.70 | 3.51 |
| Canada | 3.34 | 3.19 | |
| Australia | 3.53 | 3.35 | |
| UK | 3.59 | 3.43 | |
| Misinformation | USA | 3.53 | 3.35 |
| Canada | 3.06 | 2.92 | |
| Australia | 3.33 | 3.13 | |
| UK | 3.42 | 3.23 | |
Discussion
This study examined how young adults develop trust in nutrition influencers and their advice on TikTok, focusing on the effects of influencer qualification and information accuracy. Trust is long-established as central to the patient–healthcare professional relationship [36] and is shaped by expectations of competence, benevolence, integrity, and openness [37–39]. Today, consumers navigate a broader set of nutrition information sources that includes healthcare professionals and official guidelines alongside culinary practices, food debates, lifestyle choices, and a high volume of online recommendations from commercial actors and social media influencers [40–42]. This context is important because nutrition advice from non-credentialed influencers is often inaccurate or of suboptimal quality [13], yet social media signals and algorithmic recommendations can still promote trust and uptake. Consumers’ expectations of lower accuracy in social media contexts [43, 44] and limited effort in processing information [45] may further increase the likelihood of consuming misinformation, and recent evidence suggests health misinformation may be harder to detect than misinformation in other domains, such as political news [46]. Against this backdrop, we tested how qualification and information accuracy shape trust amongst a sample of 1,016 young adult TikTok users across the USA, United Kingdom (UK), Australia, and Canada.
Principal findings
Overall, trust in the influencer was higher across all conditions compared to trust in the message. This suggests that while respondents may trust an unknown influencer, they are generally less trusting of their advice. Additionally, trust in both the influencer and the message was higher when the nutritional advice was accurate. Two findings are central. First, influencer qualification had a positive effect on trust in the influencer and trust in the message. This indicates that disclosed credentials can operate as a trust cue in short-form nutrition content and suggests that simple credential disclosures—for example, stating qualifications in profiles and at the start of posts—may increase trust in both the communicator and the advice. Second, trust in the influencer and trust in the message were higher when the information shared was accurate. This pattern is consistent with prior work showing that users value research-based nutrition advice [5] and that authentic information can be perceived as more credible and more engaging, including when promoted by influencers [25, 29]. At the same time, the difficulty users have in identifying accurate nutrition information in some contexts [5] underscores why testing these effects is important.
Post hoc analyses suggested cross-national variation in trust, with lower levels of trust in both the influencer and the message among Canadian respondents. These results indicate that trust responses may not be uniform across English-speaking countries. Because these analyses were post hoc, they should be interpreted cautiously and treated as exploratory. Cultural factors have been shown to influence trust in online environments [47], and differences in collectivism and social harmony values, trust in institutions [48–50], and digital or health literacy could plausibly contribute to these patterns, though future work is needed to directly test these explanations.
Interpretation and contributions
This study contributes to research at the intersection of information behaviour, trust, health, and social media by providing experimental evidence that two cues, qualification and accuracy, shape trust in both the source and the message. Information Behaviour theories suggest that, under conditions of abundance, individuals may rely on information that meets minimum thresholds for anxiety avoidance, convenience, and ease [51, 52]. TDT posits that people often default to presuming honesty in communication [27], and that social media users may tend to believe content they encounter and may share content they find compelling or meaningful [53]. Our findings are consistent with these perspectives. Disclosed qualifications appear to function as an expertise or role-based trust cue that can elevate trust, and accuracy cues aligned with guidance from established institutions further elevate trust in both the influencer and the advice. However, these cues are potentially double-edged. The same mechanisms that increase trust when information is accurate may also increase misplaced trust when content is inaccurate or commercially biased, reinforcing the importance of accuracy-related signals and guidelines in nutrition communication.
Our findings also connect to work on nudging in health contexts. Nudging is commonly conceptualised as guiding decisions or behaviours without restricting choice [54]. Within the broader framework of cognitive, affective, and behavioural nudges [55], this experiment manipulated qualification and accuracy cues to shape perceptions of trust through expertise signalling [56] and alignment with WHO guidance. The results complement prior research on social media relevant nudges, including endorsement-related effects on food choice [57], follower-based nudges for healthier eating [58], and accuracy nudges to mitigate misinformation influence [59]. In this context, qualification disclosure and guideline-consistent advice may operate as practical trust cues that can be leveraged to support accurate nutrition communication and potentially counter misinformation.
Finally, this study provides empirical evidence in a platform context where experimental research remains limited. While substantial research examines food, diet, and nutrition influencers on platforms such as YouTube, Instagram, Twitter, and Facebook [22, 60, 61], experimental evidence on TikTok, particularly multi-country evidence, remains comparatively scarce. This study addresses that gap and highlights the value of considering cross-national differences rather than assuming uniform responses across similar linguistic regions or age cohorts [58].
Implications for practice
Promoting evidence-based nutrition information is important for addressing malnutrition, overweight, and obesity, and professional bodies increasingly encourage healthcare professionals to use social media for advocacy, knowledge sharing, and countering misinformation [62, 63]. Three practical implications follow from our findings. First, it is important that credentialed nutrition professionals engage on platforms used by young audiences, including TikTok. Second, disclosing credentials clearly and consistently may increase trust in both the communicator and the message, supporting the formation of trust among new or unfamiliar audiences. Third, accuracy cues, such as referencing widely accepted guidelines and evidence-based consensus, may increase trust and could complement strategies designed to counter misinformation, aligning with intervention approaches such as accuracy nudges [59]. More broadly, collaborations between credentialed professionals and high-reach non-credentialed influencers may help combine reach with expertise, however, such partnerships require careful governance and transparency, as commercial incentives or conflicting endorsements may undermine public health messaging.
Recognising potential cross-national differences in trust profiles may also matter for message design. For example, campaigns targeting Canadian audiences may benefit from emphasising trusted public institutions and guideline sources, consistent with evidence suggesting higher institutional trust in Canada [49, 50]. However, given the exploratory nature of the post hoc findings, targeted tailoring should be evaluated empirically.
Platform governance may also shape how nutrition information is encountered and trusted. Social media platforms often maintain community guidelines related to health content, yet implementation is challenging at scale. Given that qualifications can function as trust cues, professional associations and healthcare professionals may consider how best to promote relevant platform guidelines and reporting mechanisms. In addition, credentialed professionals could potentially serve as trusted flaggers under the European Union Digital Services Act (DSA), which encourages trusted flaggers to help identify and report illegal or harmful content more efficiently [64].
Limitations and future research directions
This study has limitations. First, it focused on trust outcomes rather than downstream behavioural intentions or behaviour –for example, sharing, following advice, or changes in dietary behaviour. Future research should examine whether qualification and accuracy cues translate into behavioural intentions and behaviours, and whether trust mediates these pathways, including engagement through eWOM [18, 19]. Second, the study examined initial trust in an unfamiliar influencer, which is valuable for understanding first impressions but does not capture how trust evolves over time. Longitudinal designs could assess how repeated exposure and additional trust cues –for example personal involvement, influence trust trajectories. Third, the stimuli focused on one nutrition trend, high protein intake beyond guidelines, and future work could test whether these effects generalise across other nutrition topics and risk profiles. Finally, the proliferation of content generated by artificial intelligence (AI) may alter both exposure and trust dynamics in social media environments. TikTok now requires the labelling of AI-generated content, particularly where realistic images, audio or video are created or significantly altered using AI [65]. Future research could examine whether users can identify AI-generated content and whether such labels affect trust in nutrition messages.
Conclusion
Our study examines initial trust development in nutrition influencers on TikTok among a multi-country sample of young adults. Social media platforms, including TikTok, are a source of nutrition information and advice for many users, particularly younger users, but they also facilitate the rapid spread of nutrition misinformation. Responding to calls for research [13, 66], this study examines how the accuracy of nutrition advice and influencer characteristics, specifically disclosed qualification, shape trust in nutrition content on TikTok. Our findings demonstrate that influencer qualification and information accuracy each positively influence young adults’ trust in new influencers and their advice. These results suggest practical steps for credentialed professionals to strengthen trust and support efforts to counter misinformation, and they highlight priorities for future research.
Acknowledgements
None
Abbreviations
- AI
Artificial intelligence
- ANCOVA
Analysis of Covariance
- ANOVA
Analysis of Variance
- DSA
Digital services act (European Union)
- eWOM
Electronic word-of-mouth
- H1a / H1b / H2a / H2b
Hypotheses 1a, 1b, 2a, 2b
- SPSS
Statistical Package for the social sciences
- TDT
Truth default theory
- UK
United Kingdom
- USA
United States of America
- WHO
World Health Organization
Appendix A: Survey instrument
Mis- and disinformation survey: TikTok video scripts
Transcript 1: qualified, factual information (Duration 01:02)
How much protein does your body need? My name is Sarah Hall. Many people ask me, what is the necessary amount of protein for your body? As a registered dietitian, here is what I suggest. The World Health Organisation advises a protein intake of 0.8 to 1 g per kilogram of body weight, which is considered the essential minimum. It is often misunderstood that there is no cap on the amount of protein the body can take in, but this is not the case. For those who lead a more physically demanding lifestyle, the recommended intake increases to 1.1 to 1.7 g per kilogram. This means for an active person who weighs 60 kg, the daily protein requirement ranges from 66 to 102 g. To avoid complex calculations, I suggest a straightforward approach: aim for approximately 20 g of protein per meal, and possibly up to 30 g for better results, ensuring a total intake of at least 90 g per day. By focusing on protein in your diet, you’ll likely experience greater fullness and sustained satisfaction between meals. If you have any queries, feel free to reach out.
Transcript 2: qualified, misinformation (Duration 00:56)
How much protein does your body need? My name is Sarah Hall. Many people ask me, what is the necessary amount of protein for your body? As a registered dietitian, here is what I suggest. The World Health Organisation advises a protein intake from 0.8 to 1.7 g per kilogram of body weight, which is considered the essential minimum. But the reality is that there is no cap on the amount of protein the body can take in especially as we all lead physically demanding lifestyles, so we really need a lot of protein to support that. This means for someone who weighs 60 kg, your ideal daily protein intake is a minimum 102 g, and you should really shoot for between 120 and 150 g. To avoid complex calculations, I suggest a straightforward approach, aim to eat as much protein as you can at all of your meals until you feel really sated. By focusing on protein in your diet, you’ll likely experience greater fullness and sustained satisfaction between meals. If you have any queries, feel free to reach out.
Transcript 3: unqualified, factual information (Duration 01:03)
How much protein does your body need? My name is Sarah Hall. Many people ask me, what is the necessary amount of protein for your body? While I am not a registered dietitian, here is what I advise. The World Health Organisation advises a protein intake of 0.8 to 1 g per kilogram of body weight, which is considered the essential minimum. It is often misunderstood that there is no cap on the amount of protein the body can take in, but this is not the case. For those who lead a more physically demanding lifestyle, the recommended intake increases to 1.1 to 1.7 g per kilogram. This means for someone who weighs 60 kg, the daily protein requirement ranges from 66 to 102 g. To avoid complex calculations, I suggest a straightforward approach: aim for approximately 20 g of protein per meal, and possibly up to 30 g for better results, ensuring a total intake of at least 90 g per day. By focusing on protein in your diet, you’ll likely experience greater fullness and sustained satisfaction between meals. If you have any queries, feel free to reach out.
Transcript 4: unqualified, misinformation (Duration 00:56)
How much protein does your body need? My name is Sarah Hall. Many people ask me, what is the necessary amount of protein for your body? While I am not a registered dietitian, here is what I advise. The World Health Organisation advises a protein intake of 0.8 to 1.7 g per kilogram of body weight, which is considered the essential minimum. But the reality is that there is no cap on the amount of protein the body can take in especially as we all lead physically demanding lifestyles, so we really need a lot of protein to support that. This means for someone who weighs 60 kg, your ideal daily protein intake is a minimum 102 g, and you should really shoot for between 120 and 150 g. To avoid complex calculations, I suggest a straightforward approach, aim to eat as much protein as you can at all of your meals until you feel really sated. By focusing on protein in your diet, you’ll likely experience greater fullness and sustained satisfaction between meals. If you have any queries, feel free to reach out.
Transcript 5: no qualification, factual information (Duration 01:00)
How much protein does your body need? My name is Sarah Hall. Many people ask me, what is the necessary amount of protein for your body? The World Health Organisation advises a protein intake of 0.8 to 1 g per kilogram of body weight, which is considered the essential minimum. It is often misunderstood that there is no cap on the amount of protein the body can take in, but this is not the case. For those who lead a more physically demanding lifestyle, the recommended intake increases to 1.1 to 1.7 g per kilogram. This means for someone who weighs 60 kg, the daily protein requirement ranges from 66 to 102 g. To avoid complex calculations, I suggest a straightforward approach: aim for approximately 20 g of protein per meal, and possibly up to 30 g for better results, ensuring a total intake of at least 90 g per day. By focusing on protein in your diet, you’ll likely experience greater fullness and sustained satisfaction between meals. If you have any queries, feel free to reach out.
Transcript 6: no qualification, misinformation (Duration 00:54)
How much protein does your body need? My name is Sarah Hall. Many people ask me, what is the necessary amount of protein for your body? The World Health Organisation advises a protein intake of 0.8 to 1.7 g per kilogram of body weight, which is considered the essential minimum. But the reality is that there is no cap on the amount of protein the body can take in especially as we all lead physically demanding lifestyles, so we really need a lot of protein to support that. This means for someone who weighs 60 kg, your ideal daily protein intake is a minimum 102 g, and you should really shoot for between 120 and 150 g. To avoid complex calculations, I suggest a straightforward approach, aim to eat as much protein as you can at all of your meals until you feel really sated. By focusing on protein in your diet, you’ll likely experience greater fullness and sustained satisfaction between meals. If you have any queries, feel free to reach out.
Author contributions
GF, TL, LvW and CW designed the study and oversaw the analysis, GF and AE prepared the experiments, GF, TL, AE, GS collected the data, GF, GS and LvW analysed the data, GF, TL, and LvW drafted the manuscript. All authors reviewed and approved the final manuscript.
Funding
This study was commissioned and funded by MyFitnessPal. The funders had no role in the research process nor the preparation of the manuscript.
Data availability
Data can be made available upon reasonable request to the corresponding author.
Declarations
Ethics approval
This study was reviewed and approved by the Research Ethics Committee of Dublin City University (DCUREC/2023/220) and adheres to the WMA Declaration of Helsinki. All participants provided their informed consent prior to participation.
Clinical trial number
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Grace Fox, Email: grace.fox@ucf.edu.
Theo Lynn, Email: theo.lynn@dcu.ie.
References
- 1.Moorhead SA, Hazlett DE, Harrison L, Carroll JK, Irwin A, Hoving C. A new dimension of health care: systematic review of the uses, benefits, and limitations of social media for health communication. J Med Internet Res. 2013;15(4):e1933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Grabb D. Leaning into the digital age: the role of tiktok and other technologies in providing mental health information. BMJ. 2023. 10.1136/bmj.p1754. [DOI] [PubMed] [Google Scholar]
- 3.Bursztynsky J. (2021, September 27). TikTok says 1 billion people use the app each month. CNBC. https://www.cnbc.com/2021/09/27/tiktok-reaches-1-billion-monthly-users.html
- 4.Mitchell-Wolf E. TikTok is winning over Gen Z searchers. Insider Intelligence. https://www.insiderintelligence.com/content/tiktok-winning-over-gen-z-searchers 2022.
- 5.Kreft M, Smith B, Hopwood D, Blaauw R. The use of social media as a source of nutrition information. S Afr J Clin Nutr. 2023;36(4):162–8. [Google Scholar]
- 6.Ketchum. Gen Z’S relationship with food ? It’s complicated. https://www.ketchum.com/gen-z-food-report/download-report-thank-you/. (accessed 25 July 2024) 2023
- 7.Minadeo M, Pope L. Weight-normative messaging predominates on tiktok-a qualitative content analysis. PLoS ONE. 2022;17(11):e0267997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sinapuelas IC, Ho FN. Information exchange in social networks for health care. J Consum Mark. 2019;36(5):692–702. [Google Scholar]
- 9.Bridgman A, Merkley E, Loewen PJ, Owen T, Ruths D, Teichmann L, Zhilin O, The causes and consequences of COVID-19 misperceptions: Understanding the role of news and social media. Harvard Kennedy School Misinformation Review, 1. https://misinforeview.hks.harvard.edu/article/the-causes-and-consequences-of-covid-19-misperceptions-understanding-the-role-of-news-and-social-media/ (accessed on 21 June, 2024) 2020.
- 10.Swire-Thompson B, Lazer D. Public health and online misinformation: challenges and recommendations. Annu Rev Public Health. 2020;41(1):433–51. [DOI] [PubMed] [Google Scholar]
- 11.Di Domenico G, Nunan D, Pitardi VV. Marketplaces of misinformation: a study of how vaccine misinformation is legitimised on social media. J Public Policy Mark. 2022;41(4):319–35. [Google Scholar]
- 12.Starvaggi I, Dierckman C, Lorenzo-Luaces L. Mental health misinformation on social media: review and future directions. Curr Opin Psychol. 2024;56:101738. [DOI] [PubMed] [Google Scholar]
- 13.Denniss E, Lindberg R, McNaughton SA. Quality and accuracy of online nutrition-related information: a systematic review of content analysis studies. Public Health Nutr. 2023;26(7):1345–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Karami A, Zain A, Jamal A. Unveiling the information mirage: a systematic literature review of health misinformation on social media. J Public Health. 2025;19:1–21. [Google Scholar]
- 15.MacFarlane D, Hurlstone MJ, Ecker UK. Protecting consumers from fraudulent health claims: a taxonomy of psychological drivers, interventions, barriers, and treatments. Soc Sci Med. 2020;259:112790. [DOI] [PubMed] [Google Scholar]
- 16.Henry TL, Olakunle OE. The physician’s role in countering medical misinformation through advocacy. Popul Health Manag. 2024;27(3):224–6. [DOI] [PubMed] [Google Scholar]
- 17.Lou C, Yuan S. Influencer marketing: how message value and credibility affect consumer trust of branded content on social media. J Interact Advert. 2019;19(1):58–73. [Google Scholar]
- 18.Chetioui Y, Butt I, Fathani A, Lebdaoui H. Organic food and instagram health and wellbeing influencers: an emerging country’s perspective with gender as a moderator. Br Food J. 2023;125(4):1181–205. [Google Scholar]
- 19.Mishra A, Satish SM. eWOM: extant research review and future research avenues. Vikalpa. 2016;41(3):222–33. [Google Scholar]
- 20.Folkvord F, Roes E, Bevelander K. Promoting healthy foods in the new digital era on Instagram: an experimental study on the effect of a popular real versus fictitious fit influencer on brand attitude and purchase intentions. BMC Public Health. 2020;20:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Khamis S, Ang L, Welling R. Self-branding, ‘micro-celebrity’ and the rise of social media influencers. Celebr Stud. 2017;8(2):191–208. [Google Scholar]
- 22.Engel E, Gell S, Heiss R, Karsay K. Social media influencers and adolescents’ health: a scoping review of the research field. Soc Sci Med. 2024;340:116387. [DOI] [PubMed] [Google Scholar]
- 23.Senft TM. Microcelebrity and the branded self. A companion to new media dynamics. 2013:346-54.
- 24.Burke-Garcia A, Soskin HR. Scaling the idea of opinion leadership to address health misinformation: The case for “health communication AI.” J Health Commun. 2024;29(6):396–9. [DOI] [PubMed] [Google Scholar]
- 25.Wasike B. When the influencer says jump! How influencer signaling affects engagement with COVID-19 misinformation. Soc Sci Med. 2022;315:115497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Schneider-Kamp A, Takhar J. Interrogating the pill: rising distrust and the reshaping of health risk perceptions in the social media age. Soc Sci Med. 2023;331:116081. [DOI] [PubMed] [Google Scholar]
- 27.Levine TR. Truth-default theory (TDT) a theory of human deception and deception detection. J Lang Soc Psychol. 2014;33(4):378–92. [Google Scholar]
- 28.Chua AY, Banerjee S. Intentions to trust and share online health rumors: an experiment with medical professionals. Comput Hum Behav. 2018;87:1–9. [Google Scholar]
- 29.Lorenz T. Twitter labeled factual information about covid-19 as misinformation. Available at https://www.washingtonpost.com/technology/2022/08/25/twitter-factual-covid-info-labeled-misinformation/ (accessed on 11 August 2024) 2022.
- 30.Olsacher A, Bade C, Ehlers J, Fehring L. How to effectively communicate health information on social media depending on the audience’s personality traits: an experimental study in the context of organ donation in Germany. Soc Sci Med. 2023;335:116226. [DOI] [PubMed] [Google Scholar]
- 31.Bode L, Vraga EK. See something, say something: correction of global health misinformation on social media. Health Commun. 2018;33(9):1131–40. [DOI] [PubMed] [Google Scholar]
- 32.Statistica. Countries with the largest TikTok audience as of April 2024. Available online: https://www.statista.com/statistics/1299807/number-of-monthly-unique-tiktok-users/ (accessed 11 July 2024) 2024.
- 33.Ko GJ, Rhee CM, Kalantar-Zadeh K, Joshi S. The effects of high-protein diets on kidney health and longevity. J Am Soc Nephrol. 2020;31(8):1667–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hudders L, De Jans S. Gender effects in influencer marketing: an experimental study on the efficacy of endorsements by same-vs. other-gender social media influencers on instagram. Int J Advert. 2022;41(1):128–49. [Google Scholar]
- 35.Rutherford A. ANOVA and ANCOVA: a GLM approach. Hoboken, NJ: Wiley; 2011. [Google Scholar]
- 36.Fox G, Kennedy J, van der Werff L, Lynn T. ‘Trust me, I’m a doctor’: examining the influence of threat perceptions, stigma, emotion, and distrust on reproductive health data falsification. Reproductive Health: Online First Publication; 2025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Dibben MR, Morris SE, Lean ME. Situational trust and co-operative partnerships between physicians and their patients: a theoretical explanation transferable from business practice. QJM. 2000;93(1):55–61. [DOI] [PubMed] [Google Scholar]
- 38.Hall MA, Dugan E, Zheng B, Mishra AK. Trust in physicians and medical institutions: what is it, can it be measured, and does it matter? Milbank Q. 2001;79(4):613–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Mechanic D. The functions and limitations of trust in the provision of medical care. J Health Polit Policy Law. 1998;23(4):661–86. [DOI] [PubMed] [Google Scholar]
- 40.Friedman VJ, Wright CJ, Molenaar A, McCaffrey T, Brennan L, Lim MS. The use of social media as a persuasive platform to facilitate nutrition and health behavior change in young adults: web-based conversation study. J Med Internet Res. 2022;24(5):e28063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Rogers A, Wilkinson S, Downie O, Truby H. Communication of nutrition information by influencers on social media: a scoping review. Health Promot J Aust. 2022;33(3):657–76. [DOI] [PubMed] [Google Scholar]
- 42.Wandel M. Changes in consumer concerns regarding food safety and trust in nutrition experts ‘ advice. Ecol Food Nutr. 2004;43(6):443–62. [Google Scholar]
- 43.Pennycook G, Cannon TD, Rand DG. Prior exposure increases perceived accuracy of fake news. J Exp Psychol Gen. 2018;147(12):1865–80. 10.1037/xge0000465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Pennycook G, Rand DG. Fighting misinformation on social media using crowdsourced judgments of news source quality. Proc Natl Acad Sci U S A. 2019;116(7):2521–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Peng W, Lim S, Meng J. Persuasive strategies in online health misinformation: a systematic review. Inf Commun Soc. 2023;26(11):2131–48. [Google Scholar]
- 46.Luo M, Hancock JT, Markowitz DM. Credibility perceptions and detection accuracy of fake news headlines on social media: effects of truth-bias and endorsement cues. Commun Res. 2022;49(2):171–95. [Google Scholar]
- 47.Hallikainen H, Laukkanen T. National culture and consumer trust in e-commerce. Int J Inf Manage. 2018;38(1):97–106. [Google Scholar]
- 48.Hoftede G, Hofstede GJ, Minkov M. Cultures and organizations: software of the mind: intercultural cooperation and its importance for survival. McGraw-Hill; 2010.
- 49.Nevitte N. The decline of deference: Canadian value change in cross national perspective. University of Toronto Press; 1996. [Google Scholar]
- 50.Norris P. Democratic deficit: Critical citizens revisited. Cambridge University Press; 2011. [Google Scholar]
- 51.Case DO, Given LM. Looking for information: a survey of research on information seeking, needs, and behavior. Emerald: Bingley; 2016. [Google Scholar]
- 52.Connaway LS, Dickey TJ, Radford ML. If it is too convenient I’m not going after it.” convenience as a critical factor in information seeking behaviors. Libr Inf Sci Res. 2011;33(3):179–90. [Google Scholar]
- 53.Zimmerman T, Njeri M, Khader M, Allen J, Rosellini A, Eaves T. A review of truth-default theory: implications for information behavior research. Proc Assoc Inf Sci Technol. 2020;57(1):e312. [Google Scholar]
- 54.Thaler RH, Sunstein CR. Nudge: Improving decisions about health, wealth, and happiness. Penguin; 2009
- 55.Cadario R, Chandon P. Which healthy eating nudges work best? A meta-analysis of field experiments. Mark Sci. 2020;39(3):465–86. [Google Scholar]
- 56.Kramer RM, Lewicki RJ. Repairing and enhancing trust: approaches to reducing organizational trust deficits. Acad Manag Ann. 2010;4(1):245–77. [Google Scholar]
- 57.Hawkins L, Farrow C, Thomas JM. Does exposure to socially endorsed food images on social media influence food intake? Appetite. 2021;165:105424. [DOI] [PubMed] [Google Scholar]
- 58.Charry K, Tessitore T. I tweet, they follow, you eat: number of followers as nudge on social media to eat more healthily. Soc Sci Med. 2021;269:113595. [DOI] [PubMed] [Google Scholar]
- 59.Xue X, Ma H, Zhao YC, Zhu Q, Song S. Mitigating the influence of message features on health misinformation sharing intention in social media: experimental evidence for accuracy-nudge intervention. Soc Sci Med. 2024. 10.1016/j.socscimed.2024.117136. [DOI] [PubMed] [Google Scholar]
- 60.Rini L, Schouteten JJ, Faber I, Frøst MB, Perez-Cueto FJ, De Steur H. Social media and food consumer behavior: A systematic review. Trends Food Sci Technol. 2024;143:104290. [Google Scholar]
- 61.Sina E, Boakye D, Christianson L, Ahrens W, Hebestreit A. Social media and children’s and adolescents’ diets: a systematic review of the underlying social and physiological mechanisms. Adv Nutr. 2022;13(3):913–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.British Dietetic Association. The BDA and social media. Available at https://www.bda.uk.com/about-us/the-bda-and-social-media.html (accessed 11 August 2024). 2024
- 63.Dietitians Association of Australia. Dialling into the digital age: Guidance on social media for DAA members - 2nd Edition. Available at: https://member.dietitiansaustralia.org.au/Common/Uploaded%20files/DAA/Resource_Library/2020/Guidance_on_social_media_for_DAA_Members.pdf (accessed 11 August 2024) 2024.
- 64.European Commission. The Digital Services Act. Available at: https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/europe-fit-digital-age/digital-services-act_en (accessed on 11 August 2024) 2024.
- 65.TikTok. About AI-generated content. TikTok. Available online: https://support.tiktok.com/en/using-tiktok/creating-videos/ai-generated-content (accessed 10 March 2026) 2026.
- 66.Powell J, Pring T. The impact of social media influencers on health outcomes: Systematic review. Soc Sci Med. 2024;340:116472. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data can be made available upon reasonable request to the corresponding author.